After answering to this topic, I thought it would be a good idea to share the explanations about stats in a post and illustrate it with more examples.
When something is played more, the percentage of wins will be closer to 50%.
Let's say I play 20 games and get 15wins/5losses, which gives me a 75% winrate. I have 10 more wins than losses then. Now I play up to 100 games and I keep my 10 more wins than losses, which gives me 55w/45l so a 55% winrate. The more you play so the more a talent is played, the closer the winrate will be to 50%, it is a natural statistical phenomenon!
Now admitting I still win much more games, to keep the 75% winrate you would need to have 75w/25l so 50 more wins than losses, and in HotS, a single talent choice isn't enough to determine if you win or lose by itself, you pick the correct talent among 7 tiers, among 10 heroes playing, and among infinite gameplay possibilities on a variety of maps, every single game.
This is what was answered in the topic to explain why for Li-Ming's LVL7 talents Calamity had a lower winrate than Seeker in master rank (and Zei's Vengeance on one of the stat collecting platform).
We will illustrate this by interpreting other stats on HotsLogs, always master rank SL last 2 patches:
|Apocalypse||67.5% WR||83.2% picked|
|Lightning Breath||68.5% WR||16.8% picked|
There are a few ways of interpreting those statistics:
- Apocalypse suffers from being closer to 50% because it's much more popular, so normalizing that makes Apocalypse straight up better than Lightning Breath.
- Lightning Breath is more situational than Apocalypse but performs just as good in the right situations.
Either way, both of those are pretty good and Apocalypse is more popular in higher ranks, it also seems to be overall the better choice when taking the 50% normalization in account. You could find the same thing with Rehgar's Ultimates, Bloodlust being more situational and team dependant, with a slightly lower winrate and way lower pickrate.
Anduin's lvl 7 talents:
|Binding Heal||73.7% WR||12.3% picked|
|Power Word: Shield||66.1% WR||28.7% picked|
|Blessed Recovery||59.0% WR||59% picked|
This one shows perfectly the 50% normalization effect, the more a talent is picked, the closer its winrate is to 50%. The interpretation is simple, Blessed Recovery is the correct pick in the majority of cases in the current meta. When the situation deviates from the meta or changes something for Anduin himself, other talents become better for said situations. All talents are good when picked correctly.
D.Va's lvl 1 talents:
|Full Metal||73.3% WR||84.5% picked|
|Pro Moves||49.1% WR||6.6% picked|
|Liquid Cooling||71.2% WR||8.8% picked|
Full Metal is performing extremely well with a really high pickrate, Liquid Cooling isn't picked much but when it is picked it performs just as good. Pro Moves however, performs poorly compared to the other two, and while the fact it isn't picked too much could bring up the argument that the sample isn't big enough, the more likely answer is that it is bad so it isn't picked much, this is still master rank after all.
Hope you understand the 50% Winrate normalization bias now, and that it will help you look at stats and understand them better.
Source: Original link
© Post "Today’s Question – Statistical Bias, not really a question though." for game Heroes of the Storm.
Top 10 Most Anticipated Video Games of 2020
2020 will have something to satisfy classic and modern gamers alike. To be eligible for the list, the game must be confirmed for 2020, or there should be good reason to expect its release in that year. Therefore, upcoming games with a mere announcement and no discernible release date will not be included.
Top 15 NEW Games of 2020 [FIRST HALF]
2020 has a ton to look forward to...in the video gaming world. Here are fifteen games we're looking forward to in the first half of 2020.